Guide to using advanced analytics and AI in business applications
A comprehensive collection of articles, videos and more, hand-picked by our editors
Five years ago, tech entrepreneur Marc Andreessen pronounced in the pages of The Wall Street Journal that software was eating the world. Nobody wanted to make hardware anymore -- software was higher profile and more profitable. Today, you could say the same thing about what machine learning models are doing to software.
All of the heat in the tech world is still in software, but within that segment, machine learning is by far the hottest area. If your software doesn't have machine learning embedded in it, don't even bother showing up at conferences like the Consumer Electronics Show or South by Southwest Interactive. This is doubly true for enterprise software, where the ability to use machines to sharpen marketing efforts, create new products and keep manufacturing systems healthy is separating winners from losers within industries.
On the consumer side, some of the coolest products coming out are gadgets, which may sound like a curious return to hardware for tech companies, until you realize the software is the real story. Products like Amazon Echo or Google Home are only interesting because of the machine learning models that power them. These algorithms parse user voice queries, remember preferences and schedules, and learn to improve interactions with humans over time.
Then there are the most popular pure software plays, like Facebook and Google, which leverage machine learning at deep levels to customize their offerings.
On the enterprise side, the trend of machine learning modifying Andreessen's scenario and eating software is even more apparent. Last month, Tableau Software announced plans to roll out several new features in its popular data visualization software that take advantage of machine learning. Earlier this year, leading customer relationship management vendor Salesforce unveiled Einstein, a set of artificial intelligence (AI) functions that learns what's important to a user and makes it easier to find relevant data.
Probably the most obvious sign of machine learning eating the software industry is the platform war we're seeing involving some of the biggest players in tech that are looking to become the de facto machine learning and artificial intelligence development environment.
At the recent AWS re:Invent 2016 conference, Amazon Web Services announced it would give developers access to the same machine learning models and artificial intelligence tools that power the company's Alexa AI bot. This is similar to how IBM has opened up Watson to allow developers to build apps on its cognitive engine. Google, Microsoft, Facebook and Apple are also active in this area.
As IBM's CEO Ginni Rometty said at this year's World of Watson conference, "We want Watson to be the platform for business." IBM still earns much of its revenue from more traditional technologies, like cloud, but it has staked its entire reputation on Watson's machine learning capabilities.
Whichever tech company emerges from this platform war as the leader could have the opportunity to define the next several decades of computing, much the way Microsoft did with Windows starting in the 1990s.
As we head into 2017, look for this trend to continue.
This month, we already saw ride-sharing company Uber acquire Geometric Intelligence, an AI research company whose team and technology will form the core of Uber's new AI Labs division. The unit will look for ways to implement AI and machine learning to improve trip routing, order delivery and self-driving car technology, among other things.
It's no coincidence that a software company whose main line of business has nothing to do with machine learning would look to inject machine learning into the core of its operations. This is the new standard as machine learning continues to eat all software.
How eBay uses machine learning algorithms to translate product listings
Self-driving cars should be interesting test case for machine learning models
Chatbot development gets assist from ready-to-use machine learning tools